145791 - ECMA conformance: Function.prototype.apply(), Function.prototype.call()

Status: VERIFIED FIXED

(Rhino Graveyard :: Core, defect)

Description

[Phil Schwartau]

Function.prototype.apply has a FormalParameterList of length two:

          Function.prototype.apply (thisArg, argArray)

Accordingly, Function.prototype.apply.length should return 2.
This is explicitly stated in Section 15.3.4.3 of ECMA-262 Edition 3.
See http://www.mozilla.org/js/language/ for a reference.

However, Rhino currently returns 1 instead of 2.
This was reported by [e-mail].

Comment 1

[Phil Schwartau]

This bug is the Rhino version of bug 145779 against SpiderMonkey.
The corresponding testcase for this is:

          mozilla/js/tests/ecma_3/Function/regress-145779.js

Comment 2

[Igor Bukanov]

Attachment: Extending ecma_3/Function/regress-145779.js to cover other Ecma deviations

There are other problems with apply/call non-conformance.  ECMA 15.3.4.3
requires for apply to behave in the same way when passing undefined as first or
second argument as in case of passing null:

<i>If  thisArg is null or undefined, the called function is passed the global
object as the this value. Otherwise, the called function is passed
ToObject(thisArg) as the this value. 

If argArray is null or undefined, the called function is passed no arguments.
Otherwise, if argArray is neither an array nor an arguments object (see
10.1.8), a TypeError exception is thrown. If argArray is either an array or an
arguments object, the function is passed the (ToUint32(argArray.length))
arguments argArray[0], argArray[1], ...,
argArray[ToUint32(argArray.length)&#65533;1].</i>

Similarly ECMA 15.3.4.4 states for call:

<i>If  thisArg is null or undefined, the called function is passed the global
object as the this value. Otherwise, the called function is passed
ToObject(thisArg) as the this value.</i>

Rhino does not behave as required for many cases of calling apply/call with
null or undefined but I think there is no point in creating additional bug
reports for this so the the patch extends ecma_3/Function/regress-145779.js to
check against the cited text.

Comment 3

[Igor Bukanov]

Attachment: Fixes to org/mozilla/javascript/baseFunction.java

The patch makes apply.length to return 2 and addresses apply(undefined) etc.
issues.

Comment 4

[Phil Schwartau]

Igor: thank you for such thorough and precise tests! Following
prevailing custom in the JS testsuite, I have CVS-deleted

  mozilla/js/tests/ecma_3/Function/regress-145779.js

and added Igor's tests in the following two testcases:

  mozilla/js/tests/ecma_3/Function/15.3.4.3-1.js
  mozilla/js/tests/ecma_3/Function/15.3.4.4-1.js

which test ECMA-262 Edition 3

  Section 15.3.4.3 (Function.prototype.apply)
  Section 15.3.4.4 (Function.prototype.call)

I can confirm that before Igor's patch, Rhino generated exceptions
on both these testcases, but with Igor's patch, both tests pass -

Comment 5

[Igor Bukanov]

I commited the patch

Comment 6

[Phil Schwartau]

Igor, thanks! Marking Fixed, then -

Comment 7

[Phil Schwartau]

And marking Verified FIXED. Both testcases now pass in Rhino:

          mozilla/js/tests/ecma_3/Function/15.3.4.3-1.js
          mozilla/js/tests/ecma_3/Function/15.3.4.4-1.js

And that holds for both the compiled and interpreted modes of Rhino.

Comment 8

[Igor Bukanov]

Targeting as resolved against 1.5R4